| import re |
| from datasets import load_dataset, Dataset, DatasetDict |
|
|
|
|
| def main(): |
| for config in ['high', 'middle']: |
| datasets = load_dataset('ehovy/race', config) |
|
|
| _datasets = {} |
| for split in ['train', 'validation', 'test']: |
| data = [] |
| for example in datasets[split]: |
| context = str(example['article']) |
| question = str(example['question']) |
| choices = example['options'] |
| answer = example['answer'] |
| answer_index = ord(answer) - ord('A') |
| data.append({ |
| 'context': context.strip(), |
| 'question': question.strip(), |
| 'choices': choices, |
| 'answer': answer.strip(), |
| 'answer_index': answer_index |
| }) |
| dataset = Dataset.from_list(data) |
| _datasets[split] = dataset |
|
|
| datasets = DatasetDict({ |
| 'train': _datasets['train'], |
| 'validation': _datasets['validation'], |
| 'test': _datasets['test'], |
| }) |
| datasets.push_to_hub('extraordinarylab/race', config) |
|
|
|
|
| if __name__ == '__main__': |
| main() |
